Engagement Profitability Agent
Computes per-engagement margin, utilisation, burn against the SOW baseline, and revenue leakage with evidence, feeding comparables back to estimation and proof points to case studies. Never rendered client-facing.
How it works
Ask it like
Model & guardrails
- Runs on MYAM shared · open 7B (Qwen) by default — you can point it at your own model.
- Autonomy: Drafts and acts on request, human approves before anything leaves. Nothing leaves without a human’s approval.
- Data: Restricted — tightest controls. Provenance is attached to every output.
Outcome
Fills in as real usage proves it — from a Target we set, to Eval-measured (passes our test suite), to Platform-measured (across ≥20 accounts), to Verified (customer outcomes).
Also tracked: Margin figure accuracy · Client-facing render path = none
Works well with
Accounts Payable Agent
For finance teams
Parses invoices, performs deterministic three-way matching, classifies discrepancies, runs duplicate and fraud checks including vendor bank-change verification, and prepares payment proposals for human approval. Never pays.
Moves: Duplicate-payment detection recall
Accounts Receivable Agent
For finance teams
Computes aging, scores payment risk, drafts policy-compliant dunning sequences, detects disputes and excludes them automatically, reconciles receipts, and reports DSO trend. No threats or legal assertions.
Moves: Days sales outstanding
Finance Compliance Agent
For finance teams
Builds rule checklists with jurisdiction and effective date, flags gaps, summarises regulatory changes, and assembles documentation. Explicitly not legal or tax advice; abstains where jurisdiction is unclear.
Moves: Cited checklist coverage